Self-testing under uncertainty in African health systems.

Background: Decentralised diagnostic strategies may address recognised diagnostic gaps in resource constrained settings such as African health systems through improving access to diagnostics. Self-testing for viral pathogens has emerged as a technical tool through which to operationalise diagnostic decentralisation. However, the expansion of self-testing as a decentralised diagnostic strategy across HIV, hepatitis C (HCV) and epidemic potential pathogens has raised normative, operational and epidemiological uncertainties that may condition its potential value. This thesis examined self-testing as a problem of prioritisation under uncertainty in African health systems; it interrogated whether and how self-testing could be defensibly prioritised as a decentralised diagnostic strategy. Methods and results: First, an ethical analysis was conducted to address the normative uncertainty of whether dominant ethical frameworks sufficiently accounted for the ethical tensions raised by decentralised implementation models of HIV self-testing. This used the case-example of antenatal care distribution by pregnant women to their male partners in Malawi. In decentralising diagnostics into the context of people’s relationships, existing bioethical frameworks may no longer be sufficient; ethical reasoning requires interpreting universal principles through locally situated concepts of personhood. Second, Bayesian survival analysis was used to examine the operational uncertainty of cascade outcomes and disengagement from care with self-testing for a novel disease (HCV) in programme environments of differing maturity. A multi-country cohort of 1,341 participants self-testing positive for HCV at implementation project sites in Nigeria (n=226), Cameroon (n=117) and South Africa (n=998) was analysed. Nigeria and Cameroon had lower disengagement than South Africa, with facility-based sites having lower disengagement than community-based provision. Very high disengagement in South Africa was likely related to recently established centralised treatment initiation sites providing care for people who inject drugs. Third, a simple deterministic infectious disease model was used to explore self-testing for epidemic mitigation and mortality reduction. Mitigation potential was generally modest, likely demanding operationally unrealistic self-test coverage. However, self-testing could play an incremental role in mortality reductions for pathogens with high case-fatality ratios, e.g., Ebola-archetypes. Post-test adherence to isolation measures had greater influence on transmission reduction than test sensitivity or specificity. Contribution: Self-testing emerged as a socio-technical intervention that relocates diagnostic authority from institutions into domestic and community spaces. Its defensible prioritisation as a decentralised diagnostic strategy depended on the epidemiological conditions under which it was deployed, the social and system worlds into which diagnostic authority was relocated, and whether rational use had been negotiated under conditions of uncertainty. Self-testing appeared most defensible where prolonged infectious periods and limited selfcure permitted increased opportunity for diagnostic intervention (HIV), especially if linked to curative treatment (HCV), or where transmission occurred with pathogens with a high Case Fatality Ratio (CFR) through identifiable and potentially saturable networks (Ebola). However, whether these gains were realised depends on governance arrangements, behavioural assumptions underlying post-test action, and integration into existing health systems. As decentralised diagnostics continue to expand through multiplex and AI-mediated technologies, questions of rational use, governance, and health system integration will become increasingly important in avoiding fragmentation of African health systems.